See every flaw. Capture only truth. 

Avraham Linoy, Yunon Zruk, Boaz Atiya

ArtefactSense is a system designed to detect, classify, and quantify visual distortions introduced during image acquisition. It identifies a wide range of artifacts—including blur, motion smear, moiré patterns, sensor noise, compression defects, lens aberrations, rolling-shutter deformation, and exposure-related anomalies. Using a combination of deep image-quality models and distortion-specific classifiers, the system analyzes incoming images to determine both the type and severity of distortion.  

HIT-Samsung Collaboration

Measuring the truth behind every reconstructed pixel. 

Tomer Portal, Yehonatan Ravoach, Almog Sason, Dvir Manos, Inbal Bolshinsky

InterpQual SensorCheck is a system for evaluating the quality of pixel interpolation used to compensate for dead or malfunctioning sensor pixels. Instead of merely detecting defective regions, it analyzes the fidelity of the reconstructed pixel values, identifying subtle deviations such as inconsistent gradients, edge distortions, color shifts, and local texture mismatches arising from imperfect interpolation algorithms. 

HIT-Samsung Collaboration